Hadoop mapreduse debugging environment in Eclipse, hadoopmapreduse

Hadoop mapreduse debugging environment in Eclipse, hadoopmapreduse

After the environment is set up, debug the mapreduse program. However, an error occurs constantly. I hate setting in my OS environment variables, including linux. You usually like to set the environment variables in the script of the startup program, so that the script can have an appropriate environment variable value in its own running environment. In Eclipse, I expected to set the value of the hadoop_home variable. After the setting, no matter how you debug it. All errors:

Exception in thread "main" java. lang. UnsatisfiedLinkError: org. apache. hadoop. io. nativeio. NativeIO $ Windows. access0 (Ljava/lang/String; I) Z

Baidu has encountered such problems. Is the hadoopinstallation directory of windows. dynamic libraries such as hadoop.dlland winutils.exe are missing in the bindirectory. However, I have already stored these files in the previous step, but it is still unsuccessful. After several days of searching, we finally found that the bin directory must be added to the path during the normal installation of hadoop. Therefore, add the path settings in the Eclipse debugging environment. Debugging successful.

Not to mention, pay attention to the figure:

Select the Debug configuration item and set the environment parameters.

OK. debugging is normal.

After doing so, if you want to package and release a program, you 'd better remember to configure the above environment variables in the script, otherwise all kinds of exceptional errors will hit you !!

Who can help me analyze the red debugging information displayed in eclipse after the hadoop program is executed in eclipse?

Check in localhost: 50030 and localhost: 50070. It is best to run with commands. eclipse is just a development tool.

Hello, I also learned about hadoop. I just got started and asked how to use eclipse to debug hadoop programs. For example, I wrote several java files.

1. One is to deploy the test while developing, that is, to separate the two, the efficiency will be slower.
2. The second one is what you said. If the two are together, you need to install eclipse in the operating system and then install a map/reduce plug-in to deploy it while developing.
Therefore, your problem can be solved by installing a map/reduce plug-in.
Search for details.